Accuracy and prognostic value of computer-assisted (IVOS) sperm morphology evaluations
Objective: We systematically investigated the critical aspects determining the performance and in vitro predictive value of the Hamilton Thorne Research semen analyzer (IVOS, dimension system). Design: Repeat normal sperm morphology evaluations were performed. Semen samples to be used in our gamete intrafallopian tube transfer (GIFT) program were prospectively evaluated. Setting: Reproductive Biology Unit. Subjects: Liquefied semen samples from patients attending our Andrology and assisted reproduction laboratories. Interventions: Liquified semen samples were washed once and stained using Diff-Quik stain. The standard Tygerberg GIFT procedure was followed with all supernumerary oocytes inseminated and cultures. Outcome measures: Normal sperm morphology outcomes were produced using the IVOS, according to the strict criteria. The repeat measures performed were, cell- cell, intra-slide, inter-slide and inter-laboratory. Fertilization in vitro and pregnancy outcomes were recorded. Results: The predictive probabilities for an abnormal or a normal cell given two prior abnormal or two prior normal cell outcomes was 95% and 94% respectively. The average coefficients of variation obtained for the intra-slide trial was 9.73% and for the inter- slide trial it was 15.39%. The average inter-laboratory coefficients of variation obtained ranged between 16.3% and 23.1%. The logistic regression model showed that the normal sperm morphology outcomes recorded were significantly associated with fertilization in vitro and pregnancy. Analyzing the data across the 5% normal sperm morphology cut-point; fertilization rates of 39.4% (?5%) and 62.9% (>5%) and pregnancy rates of 15.2% (? 5%) and 37.4% (>5%) were obtained. Conclusion: The IVOS used produced a level of repeatability, precision and accuracy acceptable for routine application in assisted reproduction setting.